Things That Can Go Wrong

• Don't run CODA (version 1.4) twice from a cpu.
• RPC timeout of readout controller (ROC), event builder (EB), or other component. Reset CODA: Move mouse to top left of RunControl menu, click on ``file" and choose ``reset". If a ROC is persistently absent, reset the crate, wait about 3 minutes for the reboot, then reset CODA, try again. If that doesn't work call expert, e.g. Bob Michaels.
• Resetting crates: Press for 3 sec the green buttons labelled ``Trigger Supervisor" or ``Fastbus FSCC" corresponding to the crate you want to reset. The buttons are in the computer room in the third rack from the door.
• If dataspy or dhist stop working: Stop the run, reset CODA (see above), cleanup the DD system by first typing ``kdplot" and answering ``y", then type ``dd_cleanup". Then run again.
• If the disk partition fills up (it might happen on a very long run), the ``Disk Mon" window goes red, also CODA complains. Stop the run, start a new one. Disk partition switches automatically.
• Loss of electrical power. Our computer system is on UPS power which will last about 10 minutes, and which automatically and safely shuts down the computers after 2 minutes using a daemon that senses the UPS status. However, it is wise not to rely on this and to shutdown the computers using ``cd / ; /etc/shutdown -h now" from adaq account.